A New Method for Optimization: Optimal Control

This paper focuses on a new method based on the optimal control theory to tackle the optimization problem. We transform the optimization problem into an optimal control problem, composed of a discrete-time linear time-invariant system and a cost function related to the objective function. We propose a new algorithm for obtaining the numerical solution to the optimal control problem, which uses the relation between the optimal controller of every two adjacent instants. By finding the zeros of the function associated with the initial control, the new algorithm decreases computation time effectively. The state trajectory of the system rapidly approaches the local minimum point of the original optimization problem without oscillation. Finally, we use numerical examples to evaluate our new algorithm.